🤖 Retrieval

To retrieve the most relevant knowledge passages for a given query from a data source:

results = client.retrieve(query='Where is Boston?', top_k=5, sources=['City'])

To use our intelligent Retrieval Agent for retrieval:

results = client.retrieve(query='YOUR_QUERY', top_k=5, sources=['City'], agent='on')

Enabling the retrieval agent improves search results by breaking down a complex query into multiple subqueries.

To use our Metadata Agent for retrieval:

results = client.retrieve(query='YOUR_QUERY', top_k=5, sources=['City'], metadata='on')

Turning on the metadata agent improves search results through precise metadata filtering.

To list all embedding models available to use for retrieval:

results = client.list_embedding_models()

Please find the list of available embedding models here (opens in a new tab).

More embedding models are coming soon.

To add an OpenAI key to the user account:

client.add_openai_key(key='YOUR_OPENAI_API_KEY')

To remove the added key:

client.remove_openai_key()